The present invention relates to gradient coils, for example, those used in nuclear magnetic resonance (NMR) imaging systems. Gradient coils can be used to produce linear fields, as part of a process for producing images of a target object.
Nuclear magnetic resonance imaging is one of the most versatile and fastest growing modalities in medical imaging. Since the discovery by Dr. Raymond Damadian in the early 1970s that nuclear magnetic resonance techniques can be used to scan the human body to yield useful diagnostic information, medical NMR imaging devices have been developed for obtaining NMR images of the internal structures of patients. Subsequently, much effort has been expended to improve and refine the techniques used for obtaining NMR images, as well as to determine the diagnostic utility of NMR images. As a result, NMR imaging, or magnetic resonance imaging (MRI), as it is sometimes known, has today proven to be an extremely useful tool in the medical community for the purposes of detecting and diagnosing abnormalities within the body.
Conventional magnetic resonance imaging techniques generally utilize pulsed magnetic field gradients to spatially encode NMR signals from various portions of an anatomical region of interest. The pulsed magnetic field gradients, together with radio frequency excitation of the nuclear spins and acquisition of signal information, are commonly referred to as a pulse sequence.
The basic science behind NMR imaging is well known. Pulsed current through a set of conductors will produce a magnetic field external to the conductors; the magnetic field generally has the same time course of development as the current in the conductors. The conductors may be distributed in space to produce three orthogonal gradients X, Y, and Z. Each of the gradients can be independently pulsed by a separate time-dependent current waveform.
In order to construct images from the collection of NMR signals, conventional NMR imaging equipment generally utilizes magnetic field gradients for selecting a particular slice or plane of the object to be imaged and for encoding spatial information into the NMR signals. For example, one conventional technique involves subjecting an object to a continuous static homogenous field extending along a first direction, and to sets of sequences of orthogonal magnetic field gradients. Each set of orthogonal magnetic field gradient sequences generates a magnetic field component in the same direction as the static field, but the sequences have strengths that vary along the direction of the gradients.
Generally, the NMR phenomenon occurs in atomic nuclei having an odd number of protons and/or neutrons. Due to the spins of the protons and neutrons, each such nucleus exhibits a magnetic moment. As a result, when a sample composed of such nuclei is placed in the homogeneous magnetic field, a greater number of nuclear magnetic moments align with the direction of the magnetic field to produce a net macroscopic magnetization in the direction of the field. Under the influence of the magnetic field, the magnetic moments precess about the axis of the field at a frequency that is dependent upon the strength of the applied magnetic field and on the characteristics of the nuclei. The angular precession frequency, co, also referred to as the Larmor frequency, is given by the equation xcfx89=xcex3B, where xcex3 is the gyro-magnetic ratio (which is a constant for each particular atomic nucleus) and B is the strength of the magnetic field acting upon the nuclear spins.
A typical imaging procedure involves the use of three orthogonal magnetic field gradients, X, Y, and Z, which are pulsed coordinately. For example, the Z gradient is pulsed on for two brief time periods. A 90-degree radio frequency pulse in the first time period and a 180-degree; radio frequency pulse in the second time period are used to select a slice of the anatomy of interest, and to induce the nuclear spin system within that slice to generate an NMR signal. Once the slice is selected by the Z gradient, the two remaining orthogonal gradients are used to confer spatial encoding on the NMR signal in the two orthogonal directions. For example, the Y gradient will encode on the basis of phase advances imparted on a series of signal responses by using a pulsed gradient waveform of progressively increasing area. The X gradient, which is pulsed on during the signal collection period, will frequency-encode the NMR signal in the third orthogonal direction.
When the excitation pulse is stopped, the nuclear spins tend to slowly realign or relax back to the equilibrium position. At this time, the spins emit an NMR signal, which can be detected with an RF receiver coil (which can be, and often is, the same coil as that used with the transmitter). The emitted NMR signal is dependent on three basic parameters, namely, the density of the excited nuclei, the spin-lattice (longitudinal) relaxation time (T1), and the spin-spin (transverse) relaxation time (T2). The latter two parameters are both exponential time constants that characterize the rate of return to equilibrium of the longitudinal and transverse magnetization components following the application of the perturbing RF pulse. These NMR parameters of spin density, T1, and T2 are related to the atomic nuclei subjected to the NMR phenomenon.
The NMR signal is processed to yield images that give a representation of the anatomical features in the selected slice, as well as provide soft tissue contrast. NMR signals can be processed using various algorithms, depending upon the precise nature of the data acquisition procedure. However, all of these methods rely on the ability to spatially encode the signal information by making use of the magnetic field gradients, which are time modulated and sequentially pulsed in various modes to effectuate the desired result.
For example, the received NMR signals can be transformed by utilizing, for example, conventional two-dimensional Fourier transform techniques. The magnetic field and phase-encoding magnetic field gradients encode spatial information into the collection of NMR signals so that two-dimensional images of the NMR signals in the selected plane can be:constructed. During the scanning sequence, the various magnetic field gradients are repeatedly switched on and off at the desired intervals.
Gradients are often designed using some sort of a target field approach. That is, a target field is determined based on parameters of the imaging task, such as the dimensions and composition of the imaging object. Then a gradient coil is constructed based on the current density required to produce the target field. A current density function can be represented by a truncated generalized Fourier series in two dimensions. An orthogonal set of functions, such as sines and cosines or Bessel functions, can be used. The coefficients of the generalized Fourier series are computed so as to minimize error, that is, to minimize the difference between the field produced by this current density and the ideal target field. The gradient coil is then constructed by designing wire paths that most closely approximate the current density that produces the minimum error, when current flows through the wire paths.
MRI systems, such as vertical field MRI systems, usually have round magnet poles. However, rectangular poles can be useful for magnets used as a part of a system to image a patient during surgery or when a very long field of view in one direction is desired. The method for determining gradient layout described above works well for round poles, but is less effective for a design utilizing rectangular poles, which conventionally would use straight wires to carry the field-producing current. These straight wires must have a return for the current, and conventionally yield a good gradient only if the loops can be closed very far from the region of interest, approximating an infinite filamentary current. Thus, improved gradient coils for rectangular magnet poles would be advantageous for use in MRI systems.
The present invention provides gradient, coils that include different types of wire paths over different zones of the magnet pole. For example, the gradient coils can combine straight wires with conventional curved wires (derived, for example, from a Fourier series), so that a linear gradient over a rectangular pole of finite dimensions is achieved. In such a case, the straight wires provide the linear gradient, and the curved wires act as current returns, but are designed to interfere only minimally with the linear gradient, and to minimize field effects outside the zones that have the gradient coils. The coefficients of the generalized Fourier series determining the current density in each zone are selected to produce the target field. The present invention is contemplated for use with both self-shielded (two-plane) and single-plane gradient coils. In the case of self-shielded gradient coils, the shield wires are derived from the primary wires and are slightly curved in the region where the primary wires are straight.
Thus, according to one aspect of the present invention, a process of providing a gradient field includes providing a pair of opposing magnets and associated magnet poles on opposite sides of an imaging volume, and dividing each magnet pole into a plurality of zones. For each zone, a target gradient field is determined, as well as a target current density that corresponds to the target gradient field. The target current density for each zone is modeled after a mathematical series, such that at least one mathematical series modeling the target current density for one corresponding zone is different than at least one other mathematical series modeling the target current density for another corresponding zone, and such that the mathematical series modeling the current densities at boundaries of adjacent zones are continuous. Gradient coils are formed from a plurality of discrete wires. The discrete wires are disposed on the respective magnet poles along wire paths corresponding to the mathematical series modeling the target current density for the respective zones. A current source is connected to the gradient coils to provide currents through the gradient coils, thereby providing the gradient field.
The mathematical series can approximate the target current densities according to predetermined design constraints. The number of terms of the mathematical series can be limited according to the design constraints. The the gradient field can be a linear gradient field in a target region of the imaging volume. The target region can be a rectangular cross-section in a plane parallel to opposing surfaces of the poles.
The mathematical series can be, for example, a truncated generalized Fourier series, which can be an orthogonal set of functions, such as those including a Bessel function, or a function including at least one sinusoidal term.
The poles can be rectangular-shaped. Dividing each magnet pole into a plurality of zones can include dividing each magnet pole into three zones. For example, each magnet pole can be disposed in a plane z=constant, with reference to a Cartesian coordinate system. The first zone can then be defined as bounded by the lines x=xc2x1a and y=xc2x1c, the second zone can be defined as bounded by the lines x=xc2x1a, y=c, and y=c+b, and the third zone can be defined as bounded by the lines x=xc2x1a, y=xe2x88x92c, and y=xe2x88x92cxe2x88x92b.
With reference to this layout, the wire paths can be designed such that the gradient coils produce a current density in the first zone such that             J      y        ⁡          (      x      )        =            ∑              n        ,        m              ⁢                  B        nm            ⁢      cos      ⁢              xe2x80x83            ⁢                        n          ⁢                      xe2x80x83                    ⁢          π          ⁢                      xe2x80x83                    ⁢          x                a            
and Jx is substantially zero, to generate a Y gradient field in the imaging volume. The gradient field component generated by the gradient coils in the first zone can be used to capture an image of an object component within the imaging volume. The discrete wires in the first zone can be disposed on wire paths that are substantially straight lines aligned in the y-direction. Further, the wire paths can be designed such that the gradient coils produce a current density in each of the second zone and the third zone such that                     J        y            ⁡              (                  x          ,          y                )              =                  ∑                  n          ,          m                    ⁢                        B          nm                ⁢        cos        ⁢                  xe2x80x83                ⁢                              n            ⁢                          xe2x80x83                        ⁢            π            ⁢                          xe2x80x83                        ⁢            x                    a                ⁢                  cos          ⁡                      [                                          (                                  m                  +                                      1                    2                                                  )                            ⁢              π              ⁢                              xe2x80x83                            ⁢                                                (                                      y                    -                    c                                    )                                b                                      ]                          ⁢                              xe2x80x83                    ⁢                      xe2x80x83                          ⁢        and                                J        x            ⁡              (                  x          ,          y                )              =                  ∑                  n          ,          m                    ⁢                        A          nm                ⁢        sin        ⁢                  xe2x80x83                ⁢                              n            ⁢                          xe2x80x83                        ⁢            π            ⁢                          xe2x80x83                        ⁢            x                    a                ⁢                  sin          ⁡                      [                                          (                                  m                  +                                      1                    2                                                  )                            ⁢              π              ⁢                              xe2x80x83                            ⁢                                                (                                      y                    -                    c                                    )                                b                                      ]                              
to generate an X gradient field or a Y gradient field in the imaging volume. In this case, the discrete wires disposed in the second and third zones can be current returns for the gradient coil components in the first zone, and can be disposed on wire paths that are curved lines in the z-plane. Further, the discrete wires forming the gradient coils can be shaped in the second zone and the third zone to produce a current density such that Jy is substantially zero for |y|xe2x89xa7b+c.
The wire paths can be designed such that the gradient coils produce a current density in the first zone such that       J    y    =            ∑      nm        ⁢                  B        nm            ⁢              sin        ⁡                  (                      n            +                          1              2                                )                    ⁢              xe2x80x83            ⁢                        π          ⁢                      xe2x80x83                    ⁢          x                a            
and Jx is substantially zero, to generate a Z gradient field in the imaging volume. In this case, the discrete wires in the first zone can be disposed on wire paths that are substantially straight lines aligned in the y-direction. Further, the discrete wires forming the gradient coils in each of the second zone and the third zone can be shaped to produce a current density such that Jy=0 for |y|xe2x89xa7b+c.
In all the above cases, the wire paths in the first zone can be substantially straight lines, and the wire paths in the second and third zones can be curved lines.
According to another aspect of the present invention, a magnetic resonance imaging system includes a pair of opposing magnets that provide a main magnetic field in an imaging volume, a pair of magnet poles, and a pair of gradient coils. The magnet poles are disposed opposing faces of respective magnets. The gradient coils are formed from discrete wires disposed along wire paths, one on each magnet pole. Each magnet pole is divided into a plurality of zones, such that each zone is defined by a shape of the wire path in that zone. The shape in each zone corresponds to a mathematical series. The mathematical series corresponding to the shapes are different in at least two of the zones.
The system can include a current source, connected to the gradient coils to provide a current through the gradient coils; thereby providing the gradient field. Each gradient coil can provide a linear gradient field in a target region of the imaging volume, and the target region can have a rectangular cross-section in a plane parallel to the poles.
The wire paths in each said zone can be shaped to correspond to a truncated generalized Fourier series. The truncated generalized Fourier series can approximate the current density in the corresponding zone, according to predetermined design constraints. The number of terms of the truncated generalized Fourier series can be limited according to the design constraints. The truncated generalized Fourier series can be an orthogonal set of functions, such as Bessel functions, or functions including at least one sinusoidal term.
The poles can be rectangular-shaped, and each magnet pole can be divided into three zones. For example, with reference to a Cartesian coordinate system, each said magnet pole can be disposed in the plane z=constant. The first zone can then be defined as bounded by the lines x=xc2x1a and y=xc2x1c, the second zone can be defined as bounded by the lines x=xc2x1a, y=c, and y=c+b, and the third zone can be defined as bounded by the lines x=xc2x1a, y=xe2x88x92c, and y=xe2x88x92cxe2x88x92b.
With reference to the foregoing exemplary layout, the wire paths can be shaped such that the gradient coils produce a current density in the first zone such that             J      y        ⁡          (      x      )        =            ∑      nm        ⁢                  B        nm            ⁢      cos      ⁢              xe2x80x83            ⁢                        n          ⁢                      xe2x80x83                    ⁢          π          ⁢                      xe2x80x83                    ⁢          x                a            
and Jx is substantially zero, to generate a Y gradient field in the imaging volume. The discrete wires in the first zone can be disposed on wire paths that are straight lines aligned in the y-direction. Further, the wire paths can be shaped such that the gradient coils produce a current density in each of the second zone and the third zone such that                     J        y            ⁡              (                  x          ,          y                )              =                  ∑                  n          ,          m                    ⁢                        B          nm                ⁢        cos        ⁢                  xe2x80x83                ⁢                              n            ⁢                          xe2x80x83                        ⁢            π            ⁢                          xe2x80x83                        ⁢            x                    a                ⁢                  cos          ⁡                      [                                          (                                  m                  +                                      1                    2                                                  )                            ⁢              π              ⁢                              xe2x80x83                            ⁢                                                (                                      y                    -                    c                                    )                                b                                      ]                          ⁢                              xe2x80x83                    ⁢                      xe2x80x83                          ⁢        and                                J        x            ⁡              (                  x          ,          y                )              =                  ∑                  n          ,          m                    ⁢                        A          nm                ⁢        sin        ⁢                  xe2x80x83                ⁢                              n            ⁢                          xe2x80x83                        ⁢            π            ⁢                          xe2x80x83                        ⁢            x                    a                ⁢                  sin          ⁡                      [                                          (                                  m                  +                                      1                    2                                                  )                            ⁢              π              ⁢                              xe2x80x83                            ⁢                                                (                                      y                    -                    c                                    )                                b                                      ]                              
to generate an X gradient field or a Y gradient field in the imaging volume. The discrete wires in the second and third zones can be disposed on wire paths that are curved lines in the z-plane. Further, the wire paths can be shaped such that the gradient coils produce a current density such that Jy=0 for |y|xe2x89xa7b+c.
Further, the wire paths can be shaped such that the gradient coils produce a current density in the first zone such that       J    y    =            ∑      nm        ⁢                  B        nm            ⁢              sin        ⁡                  (                      n            +                          1              2                                )                    ⁢                        π          ⁢                      xe2x80x83                    ⁢          x                a            
and Jx is substantially zero, to generate a Z gradient field in the imaging volume. The discrete wires in the first zone can be disposed on wire paths that are straight lines aligned in the y-direction. Thus, the discrete wires in the first zone can be disposed on wire paths that are substantially straight lines aligned in the y-direction, and the discrete wires in the second and third zones can be disposed on wire paths that are curved lines in the z-plane.